The second development is the increasing diversity of phenotypes that are being tested: these include quality of maternal parenting [26], affective state during marriage [27], risky sexual behavior [28], childhood emotionality [29], job satisfaction [30], perceived racial discrimination [31], adult unresolved attachment [32], and gaze bias [33]. A common feature of all these studies is the relatively small sample size: all except one [30] use samples of less than a thousand, and sometimes less than one hundred subjects [32]. Given the now well established main effect sizes discussed above, it seems unlikely, even allowing for the large effects observed in an interaction analysis, that any of these studies is sufficiently well powered to detect an effect.
